Skip to content

Instantly share code, notes, and snippets.

@MotionDesignStudio
Last active January 3, 2024 03:26
Show Gist options
  • Save MotionDesignStudio/9374326 to your computer and use it in GitHub Desktop.
Save MotionDesignStudio/9374326 to your computer and use it in GitHub Desktop.
Python ASCII Video And ASCII Image Creator
Example:
Video
http://www.youtube.com/watch?v=yi5H_bnZzGc
[ Usage: To convert a video file to ASCII video with no audio. ]
./ascii_movie_image_ver_1.py myvideo.mkv 4 15
“myvideo.mkv” is the name of your video file. Supported video files are limited based on your system.
“4” is the density of ASCII characters.
“15” is the font size.
Changing the density and font size changes the look and feel of the video.
[ Usage: To convert a image file to a ASCII image. ]
./ascii_movie_image_ver_1.py myimage.jpg 4 15
“myimage.jpg” is the name of your image file. Supported image files are limited based on your system.
“4” is the density of ASCII characters.
“15” is the font size.
Changing the density and font size changes the look and feel of the image.
[ How To Modify ]
The following line: font = ImageFont.truetype("/usr/share/fonts/truetype/freefont/FreeMonoBold.ttf", fontsize, encoding="unic")
Is the path to the font being utilized. If you encounter errors with the font location point this to a location where your system stores a fixed width truetype font.
Modify any lines with “120,20,20” to change the background color of the image or video.
Modify any lines with “255,255,0” to change the font color.
The color is expressed as an RGB triplet (r,g,b), each component of which can vary from 0 to 255. If all the components are at zero the result is black; if all are at maximum, the result is the brightest representable white.
External link to FFMEG:
Modify this line to your external ffmpeg source.
subprocess.call(["./ffmpeg/ffmpeg", "-i", "%d.png", "-r", frame_rate, "-y", "-c:v", "libx264", "-preset", "slow", "-crf", "18", "-c:a", "copy", "-pix_fmt", "yuv420p", "output.mkv" ])
[ Important ]
Non fixed width fonts might distort the image.
I use a static ffmpeg build because it runs faster and creates a smaller higher quality video. The ffmpeg commands I use is optimized for Youtube streaming. You can download the latest from. http://johnvansickle.com/ffmpeg/
#!/usr/bin/env python
import sys
import cv2
import subprocess
from subprocess import call
import aalib
import Image
import PIL
from PIL import ImageFont
from PIL import Image
from PIL import ImageDraw
import imghdr
import ImageOps
import numpy as np
sys.dont_write_bytecode = True
videofilename=str(sys.argv[1]).strip()
#These two variables control the resolution and font size. They are interconnected.
#img_new_magnitude
#fontsize
try:
img_new_magnitude= int (str(sys.argv[2]).strip() )
except :
img_new_magnitude=4
try:
fontsize= int (str(sys.argv[3]).strip())
except :
fontsize= 15
font = ImageFont.truetype("/usr/share/fonts/truetype/freefont/FreeMonoBold.ttf", fontsize, encoding="unic")
def create_ascii_video_2():
#Extract the image size:
image_width, image_height =calculate_image_size_from_video ()
#Extract the number of rows and length of the string in pixels which is used for the image being
#created based on the width and height of the text.
how_many_rows, new_img_width = calculate_image_size_from_video_2 ()
vc = cv2.VideoCapture(videofilename)
saved_file_name=""
frame_rate=str (vc.get(cv2.cv.CV_CAP_PROP_FPS))
#Png file name counter
c=0
if vc.isOpened():
rval , frame = vc.read()
else:
rval = False
while rval:
rval, frame = vc.read()
#Check to see if file is a valid image. I notice the last frame of the video is an empty file at times:
try:
myimage= Image.fromarray(frame)
except (TypeError, AttributeError):
print ("cannot identify image while loop in function create_ascii_video_2", myimage)
continue
aalib_screen_width= int(image_width/25.28)*img_new_magnitude
aalib_screen_height= int(image_height/41.39)*img_new_magnitude
screen = aalib.AsciiScreen(width=aalib_screen_width, height=aalib_screen_height )
myimage= Image.fromarray(frame).convert("L").resize(screen.virtual_size)
screen.put_image((0,0), myimage)
img=Image.new("RGBA", (new_img_width, how_many_rows*fontsize),(120,20,20))
draw = ImageDraw.Draw(img)
y = 0
for lines in screen.render().splitlines():
draw.text( (0,y), lines, (255,255,0),font=font )
y = y + fontsize
imagefit = ImageOps.fit(img, (image_width, image_height), Image.ANTIALIAS)
imagefit.save(str(c)+'.png', "PNG")
c = c + 1
cv2.waitKey(1)
vc.release()
cv2.destroyAllWindows()
subprocess.call(["./ffmpeg/ffmpeg", "-i", "%d.png", "-r", frame_rate, "-y", "-c:v", "libx264", "-preset", "slow", "-crf", "18", "-c:a", "copy", "-pix_fmt", "yuv420p", "output.mkv" ])
clean_up(c)
def create_ascii_image():
myimage= Image.open(videofilename)
image_width, image_height = myimage.size
aalib_screen_width= int(image_width/24.9)*img_new_magnitude
aalib_screen_height= int(image_height/41.39)*img_new_magnitude
screen = aalib.AsciiScreen(width=aalib_screen_width, height=aalib_screen_height )
myimage= Image.open(videofilename).convert("L").resize(screen.virtual_size)
screen.put_image((0,0), myimage)
y = 0
how_many_rows = len ( screen.render().splitlines() )
new_img_width, font_size = font.getsize (screen.render().splitlines()[0])
img=Image.new("RGBA", (new_img_width, how_many_rows*fontsize),(120,20,20))
draw = ImageDraw.Draw(img)
for lines in screen.render().splitlines():
draw.text( (0,y), lines, (255,255,0),font=font )
y = y + fontsize
imagefit = ImageOps.fit(img, (image_width, image_height), Image.ANTIALIAS)
imagefit.save("ascii_photo.png", "PNG")
def clean_up(howmuchtodelete):
i=0
while i< howmuchtodelete :
subprocess.call(["rm", str(i)+".png" ])
i=i+1
def complete_alert_message(the_message):
font = ImageFont.truetype("/usr/share/fonts/truetype/freefont/FreeMonoBold.ttf",40)
img=Image.new("RGBA", (200,70),(120,20,20))
draw = ImageDraw.Draw(img)
draw.text((5, 0),the_message,(255,255,0),font=font)
img.show()
def calculate_image_size_from_video ():
vc = cv2.VideoCapture(videofilename)
if vc.isOpened():
rval , frame = vc.read()
#Extract the image size:
try:
myimage= Image.fromarray(frame)
return myimage.size
except IOError:
print ("cannot identify image file", myimage)
else:
print ("In fucntion calculate_image_size_from_video I could not find the video frame size." )
def calculate_image_size_from_video_2 ():
#Extract the image size:
image_width, image_height =calculate_image_size_from_video ()
vc = cv2.VideoCapture(videofilename)
if vc.isOpened():
rval , frame = vc.read()
else:
rval = False
while rval:
rval, frame = vc.read()
#Check to see if file is a valid image. I notice the last frame of the video is an empty file at times:
try:
myimage= Image.fromarray(frame)
except (TypeError, AttributeError):
print ("cannot identify image", myimage)
continue
aalib_screen_width= int(image_width/25.28)*img_new_magnitude
aalib_screen_height= int(image_height/41.39)*img_new_magnitude
screen = aalib.AsciiScreen(width=aalib_screen_width, height=aalib_screen_height )
myimage= Image.fromarray(frame).convert("L").resize(screen.virtual_size)
screen.put_image((0,0), myimage)
how_many_rows = len ( screen.render().splitlines() )
new_img_width, font_size = font.getsize (screen.render().splitlines()[0])
cv2.waitKey(1)
vc.release()
return (how_many_rows, new_img_width )
def test_if_image_or_video ():
if imghdr.what(videofilename) != None :
print ("This is a image file.")
create_ascii_image()
else:
print ("This is probably not a image.")
create_ascii_video_2()
test_if_image_or_video ()
@hellocatfood
Copy link

I think you should explicitly state what the dependencies are. I didn't know I had to find/install python-aalib

@Kagami
Copy link

Kagami commented Oct 28, 2015

@GeekFreaker
Copy link

How I get the char sequence out without making it a png?

@unxed
Copy link

unxed commented Oct 17, 2020

how to run on recent ubuntu:

sudo apt install python3-opencv python3-autopep8 python3-aalib
autopep8 -i ascii_movie_image_ver_1.py

commented
import Image

changed
import ImageOps
to
from PIL import ImageOps

run with
python3 ascii_movie_image_ver_1.py imagefile.jpg

@elig0n
Copy link

elig0n commented Jan 16, 2021

also had to modify line 52 to
frame_rate=str (vc.get(cv2.cv2.CAP_PROP_FPS))

@MotionDesignStudio
Copy link
Author

MotionDesignStudio commented Jan 16, 2021 via email

@elig0n
Copy link

elig0n commented Jan 18, 2021

I ran it on some video files and it would only create output.mkv

@MotionDesignStudio
Copy link
Author

MotionDesignStudio commented Jan 18, 2021 via email

@elig0n
Copy link

elig0n commented Jan 18, 2021

I also had to manually set path to ffmpeg and itwould then only create red&yellow tinted videos & images
I actually thought this would play the ascii-video in terminal by terminal drawing commands

@MotionDesignStudio
Copy link
Author

MotionDesignStudio commented Jan 18, 2021 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment